HTTP Status Code Reference

How It Works

Every HTTP response starts with a three-digit code, and the first digit alone tells you which of the five classes you are dealing with. This reference covers the 46 codes a developer realistically meets, each with a one-line summary and an expanded note covering the common cause and the first action to try — filter by class, search by code or phrase, and click any row to open it.

The five classes

1xx is informational — interim signals that usually complete inside the client's connection handling. 2xx is success, with sub-flavors for created (201), accepted-but-unfinished (202) and deliberately empty (204). 3xx is redirection: go somewhere else, under stated rules. 4xx is a client error — the request as sent was unacceptable. 5xx is a server error — the request was fine, the server could not deliver. The single most useful division when debugging: 4xx means look at what you sent, 5xx means look at what they run.

Idempotency and retries

Safe to retry with backoff: 408, 425, 429 (honor Retry-After), and the 5xx family — 500, 502, 503, 504. Never worth blind retries: 4xx codes about the payload itself (400, 413, 415) and authorization failures (401, 403). Codes like 409 and 412 are retryable only after you fetch fresh state and fix the conflict — replaying the same request just fails again.

Permanent vs temporary redirects

301 and 308 are cached by clients essentially forever: use them only when the old URL will never return, and know a mistake keeps following users after you fix the server. 302, 303 and 307 are not stored — the browser re-asks the original URL next time — and differ in how the method is rewritten: 303 forces GET, 307 preserves the method, 302 keeps the method in practice but not by guarantee.

Frequently Asked Questions